ebook img

Matrix Algorithms in MATLAB PDF

468 Pages·2016·6.13 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Matrix Algorithms in MATLAB

Matrix Algorithms in ® MATLAB Matrix Algorithms in ® MATLAB Ong U. Routh AMSTERDAM • BOSTON • HEIDELBERG • LONDON NEWYORK • OXFORD • PARIS • SANDIEGO SANFRANCISCO • SINGAPORE • SYDNEY • TOKYO AcademicPressisanImprintofElsevier Publisher:JoeHayton SeniorEditorialProjectManager:KattieWashington ProjectManager:AnushaSambamoorthy Designer:MariaInesCruz AcademicPressisanimprintofElsevier 125,LondonWall,EC2Y,5AS 525BStreet,Suite1800,SanDiego,CA92101-4495,USA 50HampshireStreet,5thFloor,CambridgeMA02139,USA TheBoulevard,LangfordLane,Kidlington,OxfordOX51GB,UK Copyright©2016ElsevierInc.Allrightsreserved. Nopartofthispublicationmaybereproducedortransmittedinanyformorbyanymeans,electronicormechanical,including photocopying,recording,oranyinformationstorageandretrievalsystem,withoutpermissioninwritingfromthepublisher. Detailsonhowtoseekpermission,furtherinformationaboutthePublisher’spermissionspoliciesandourarrangementswith organizationssuchastheCopyrightClearanceCenterandtheCopyrightLicensingAgency,canbefoundatourwebsite: www.elsevier.com/permissions. ThisbookandtheindividualcontributionscontainedinitareprotectedundercopyrightbythePublisher(otherthanasmaybe notedherein). Notices Knowledgeandbestpracticeinthisfieldareconstantlychanging.Asnewresearchandexperiencebroadenourunderstanding, changesinresearchmethodsorprofessionalpractices,maybecomenecessary.Practitionersandresearchersmustalwaysrelyon theirownexperienceandknowledgeinevaluatingandusinganyinformationormethodsdescribedherein.Inusingsuch informationormethodstheyshouldbemindfuloftheirownsafetyandthesafetyofothers,includingpartiesforwhomtheyhave aprofessionalresponsibility. Tothefullestextentofthelaw,neitherthePublishernortheauthors,contributors,oreditors,assumeanyliabilityforanyinjury and/ordamagetopersonsorpropertyasamatterofproductsliability,negligenceorotherwise,orfromanyuseoroperationof anymethods,products,instructions,orideascontainedinthematerialherein. LibraryofCongressCataloging-in-PublicationData AcatalogrecordforthisbookisavailablefromtheLibraryofCongress BritishLibraryCataloging-in-PublicationData AcataloguerecordforthisbookisavailablefromtheBritishLibrary. ISBN:978-0-12-803804-8 ForinformationonallAcademicPresspublications visitourwebsiteatwww.elsevier.com ThisbookisdedicatedtomywifeKathy,andtwochildrenJeffreyandJennifer Contents List of Figures....................................................................................................xiii Preface..............................................................................................................xv License Terms....................................................................................................xix CHAPTER 1 Introduction 1 Introduction....................................................................................1 1.1 Elements of Linear Algebra................................................................1 1.1.1 Definitions...........................................................................2 1.1.2 Linear Independence and Related Concepts................................5 1.1.3 Solution of Linear Equations....................................................7 1.1.4 Solution of Eigenvalue Problem................................................9 1.2 A Brief Introduction of MATLAB........................................................15 1.3 Types of Matrices...........................................................................21 1.3.1 Square vs. Non-Square Matrices.............................................22 1.3.2 Symmetric vs. Non-Symmetric Matrices...................................22 1.3.3 Full Rank vs. Deficient Rank Matrices .....................................22 1.3.4 Singular vs. Non-Singular Matrices.........................................23 1.3.5 Orthogonal vs. Non-Orthogonal Matrices...................................24 1.3.6 Defective vs. Non-Defective Matrices.......................................25 1.3.7 Positive (semi-)Definite vs. Positive Indefinite Matrices...............28 1.3.8 Zero Structured vs. Full Matrices............................................28 1.4 Overview of Matrix Computations ......................................................36 1.5 Reordering of Sparse Matrices..........................................................46 1.6 Utility Codes.................................................................................50 CHAPTER 2 Direct Algorithms of Decompositions of Matrices by Non-Orthogonal Transformations 63 Introduction .................................................................................63 2.1 Gauss Elimination Matrix.................................................................64 2.2 LU Decomposition.........................................................................67 2.2.1 LU Decomposition by Gauss Elimination..................................67 2.2.2 LU Decomposition by Crout Procedure.....................................70 2.2.3 Driver of LU Decomposition...................................................73 2.2.4 LU Decomposition of an Upper Hessenberg Matrix.....................77 2.2.5 LU Decomposition of a Band Matrix........................................79 2.3 LDU Decomposition.......................................................................80 2.3.1 LDU Decomposition by Gauss Elimination................................81 2.3.2 LDU Decomposition by Crout Procedure...................................83 2.3.3 Driver of LDU Decomposition.................................................85 2.4 Congruent Decomposition Algorithms for Symmetric Matrices .................87 2.4.1 Reduction of Symmetric Matrix to Diagonal (LDLt).....................88 2.4.2 Cholesky Decomposition (LLt)................................................92 2.4.3 Reduction of Symmetric Matrix to Tri-Diagonal (LTLt)..................95 vii viii Contents 2.4.4 Reduction of Symmetric Matrix to Block Diagonal (LBLt)........... 101 2.4.5 Modified Cholesky Decomposition (xLLt)................................ 105 2.5 Similarity Decomposition Algorithms................................................ 110 2.5.1 Reduction of Square Matrix to Hessenberg by Gauss Elimination......................................................... 110 2.5.2 Reduction of Square Matrix to Tri-Diagonal by Gauss Elimination 114 2.5.3 Reduction of Square Matrix to Tri-Diagonal by Lanczos Procedure 118 2.6 Reduction of a Symmetric Matrix to Tri-Diagonal and Another Symmetric Matrix to Diagonal of ±1s and 0s..................................................... 122 2.6.1 Hyper Rotation and Hyper Reflection..................................... 123 2.6.2 GTJGt Decomposition by Hyperbolic Rotation or Hyperbolic Reflection........................................................................ 129 CHAPTER 3 Direct Algorithms of Decompositions of Matrices by Orthogonal Transformations 137 Introduction................................................................................ 137 3.1 Householder Reflection Matrix and Givens Rotation Matrix ................... 138 3.2 QR Decomposition....................................................................... 140 3.2.1 QR Decomposition by Householder Reflections........................ 141 3.2.2 QR Decomposition by Givens Rotations.................................. 142 3.2.3 QR Decomposition by Gram–Schmidt Orthogonalizations........... 143 3.2.4 Driver of QR Decomposition................................................. 145 3.2.5 QR Decomposition of an Upper Hessenberg Matrix................... 150 3.2.6 QR Decomposition of a Band Matrix...................................... 152 3.3 Complete Orthogonal Decomposition (QLZ) ....................................... 153 3.4 Reduction of Matrix to Bi-Diagonal.................................................. 155 3.4.1 QBZ Decomposition by Householder Reflections...................... 155 3.4.2 QBZ Decomposition by Givens Rotations................................ 159 3.4.3 QBZ Decomposition by Golub–Kahan–Lanczos Procedure .......... 162 3.5 Reduction of Square Matrix to Hessenberg by Similarity Transformations. 166 3.5.1 QHQt Decomposition by Householder Reflections..................... 167 3.5.2 QHQt Decomposition by Givens Rotations............................... 169 3.5.3 QHQt Decomposition by Arnoldi Procedure............................. 171 3.6 Reduction of Symmetric Matrix to Tri-Diagonal by Congruent Transformations........................................................................... 174 3.6.1 QTQt Decomposition by Householder Reflections..................... 174 3.6.2 QTQt Decomposition by Givens Rotations............................... 176 3.6.3 QTQt Decomposition by Lanczos Procedure ............................ 177 3.7 Reduction of a Matrix to Upper Hessenberg and Another Matrix to Upper Triangular .................................................................................. 181 3.7.1 QHRZ Decomposition by Orthogonal Transformations................ 181 3.7.2 QHRZ Decomposition by Arnoldi Procedure ............................ 184 Contents ix CHAPTER 4 Direct Algorithms of Solution of Linear Equations 187 Introduction ............................................................................... 187 4.1 Brief Introduction of Pseudo-Inverse................................................ 188 4.2 Linear Constraints to Linear Equations............................................. 191 4.2.1 When A is a Rectangular Matrix........................................... 192 4.2.2 When A is a Square Matrix.................................................. 192 4.2.3 When Bx=b Has No Solution.............................................. 198 4.3 Solution of Five Elementary Linear Equations.................................... 199 4.3.1 Linear Equations of a Zero Matrix: Ox=b.............................. 200 4.3.2 Linear Equations of a Diagonal Matrix: Dx=b......................... 200 4.3.3 Linear Equations of an Orthogonal Matrix: Qx=b.................... 201 4.3.4 Linear Equations of a Lower Triangular Matrix: Lx=b............... 202 4.3.5 Linear Equations of an Upper Triangular Matrix: Ux=b............. 204 4.4 Gauss and Gauss-Jordan Elimination Algorithms ................................ 207 4.5 Householder and Givens Elimination Algorithms................................. 216 4.6 Solution Algorithms Based on Matrix Decompositions.......................... 219 4.7 Linear Systems Arising From Interpolation........................................ 226 4.7.1 Introducing Function Interpolation........................................ 226 4.7.2 Polynomial Interpolation..................................................... 228 4.7.3 Trigonometric Interpolation.................................................. 234 CHAPTER 5 Iterative Algorithms of Solution of Linear Equations 245 Introduction ............................................................................... 245 5.1 Overview of Iterative Algorithms...................................................... 246 5.2 Stationary Iterations: Jacobi, Gauss-Seidel, and More.......................... 247 5.3 General Methodology of Non-Stationary Iterations............................... 256 5.4 Non-Stationary Iterations Applied to Symmetric Matrix........................ 259 5.4.1 CG:ConjugateGradientVisFullOrthogonalBasesofK(A,y ,m) 0 and W=V ...................................................................... 259 5.4.2 CR:ConjugateResidualVisFullOrthogonalBasesofK(A,y ,m) 0 and W=AV..................................................................... 261 5.4.3 ‘C’E: ‘Conjugate’ Error W is Full Orthogonal Bases of K(A,y ,m) 0 and V=AW..................................................................... 264 5.4.4 Numerical Examples for CG,CR, and CE................................. 266 5.5 Non-Stationary Iterations Applied to Unsymmetric Matrix..................... 270 5.5.1 FOM: Full Orthogonalization Method V is Full Orthogonal Bases of K(A,y ,m) and W=V.................................................... 270 0 5.5.2 GMRES: Generalized Minimum Residual V is Full Orthogonal Bases of K(A,y ,m) and W=AV ......................................... 272 0 5.5.3 GMERR:GeneralizedMinimumErrorWisFullOrthogonalBases of K(A′,y ,m) and V=A′W................................................ 274 0 5.5.4 DIOM: Direct Incomplete Orthogonalization Method V is Incom- plete Orthogonal Bases of K(A,y ,m) and W=? ..................... 275 0 x Contents 5.5.5 DQGMRES: Direct Quasi-Generalized Minimum Residual V is Incomplete Orthogonal Bases of K(A,y ,m) and W=?............. 278 0 5.5.6 DQGMERR:DirectQuasi-GeneralizedMinimumErrorWisIncom- plete Orthogonal Bases of K(A′,y ,m) and V=? ..................... 280 0 5.5.7 BCG:Bi-ConjugateGradientVandWareBi-OrthogonalBasesof K(A,y ,m) and K(A′,y ,m)................................................. 282 0 0 5.5.8 QMR: Quasi-Minimal Residual V is Bi-Orthogonal Bases of K(A,y ,m) to K(A′,y ,m) and W=?..................................... 284 0 0 5.5.9 QME: Quasi-Minimal Error W is Bi-Orthogonal Bases of K(A,y ,m) to K(A′,y ,m) and V=?..................................... 286 0 0 5.5.10 TFBCG:TransposedFreeBi-ConjugateGradientV=?andW=? to Make it Equivalent to BICGSTAB....................................... 287 5.5.11 TFQMR:TransposeFreeQuasi-MinimalResidualV=?andW=? to Make it Equivalent to TFQMR........................................... 291 5.5.12 Numerical Examples.......................................................... 292 5.6 Special Algorithms for Normal Equations.......................................... 294 5.6.1 Symmetrizing Transformations............................................. 295 5.6.2 Stationary Iterations for Normal Equations.............................. 296 5.6.3 Conjugate Gradient for Normal Equations............................... 300 5.6.4 RemarksonIterationAlgorithmsforConstrainedLinearEquations303 5.7 Other Important Topics................................................................. 304 5.7.1 Preconditioning Techniques................................................. 305 5.7.2 Parallel Computations........................................................ 307 5.7.3 Algebraic Multigrid Method................................................. 308 5.7.4 Domain Decomposition Method............................................ 310 CHAPTER 6 Direct Algorithms of Solution of Eigenvalue Problem 317 Introduction ............................................................................... 317 6.1 Algorithms of Solution of 2×2 Eigenvalue Problems ........................... 318 6.2 Bound Estimation of Symmetric Eigenvalue Problems ......................... 321 6.3 Power Iterations and Subspace Iterations.......................................... 326 6.4 QR Iterations.............................................................................. 327 6.4.1 Reduce A to Hessenberg Form............................................. 328 6.4.2 A QR Iteration Step for a Hessenberg Matrix........................... 328 6.4.3 Deflation......................................................................... 329 6.4.4 Shift............................................................................... 330 6.4.5 MATLAB Implementations of QR Iterations Algorithms.............. 333 6.5 Calculation of Eigenvectors by Inverse Iterations ................................ 336 6.6 Calculation of Invariant Subspace by Eigenvalue Reordering................. 340 6.7 Special Algorithms for Symmetric Tri-Diagonal Matrix.......................... 344 6.7.1 Eigenvalues by Bisection.................................................... 345 6.7.2 A Divide-and-Conquer Algorithm........................................... 347 6.8 Jacobi Iterations for Symmetric Matrix............................................. 353 6.9 Algorithms for Symmetrical Positive Definite Ax= λBx........................ 355 6.10 Algorithms for Unsymmetrical Ax= λBx........................................... 359

Description:
Matrix Algorithms in MATLAB focuses on the MATLAB code implementations of matrix algorithms. The MATLAB codes presented in the book are tested with thousands of runs of MATLAB randomly generated matrices, and the notation in the book follows the MATLAB style to ensure a smooth transition from formul
See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.